circulating cardiac troponin i levels measured by a novel...
TRANSCRIPT
Grodin, J. L. et al. (2018) Circulating cardiac troponin I levels measured by a novel
highly sensitive assay in acute decompensated heart failure: insights from the
ASCEND-HF trial. Journal of Cardiac Failure, 24(8), pp. 512-519.
(doi:10.1016/j.cardfail.2018.06.008)
There may be differences between this version and the published version. You are
advised to consult the publisher’s version if you wish to cite from it.
http://eprints.gla.ac.uk/165966/
Deposited on: 30 July 2018
Enlighten – Research publications by members of the University of Glasgow
http://eprints.gla.ac.uk
2
Circulating Cardiac Troponin I Levels Measured by a Novel Highly Sensitive Assay in
Acute Decompensated Heart Failure: Insights from the ASCEND-HF Trial
Short title: High-Sensitivity Troponin I in ADHF
Justin L Grodin MD MPH1, Javed Butler MD
2; Marco Metra MD
3, G. Michael Felker MD
4,
Adriaan A Voors MD5, John J McMurray MD
6, Paul W Armstrong MD
7, Adrian F. Hernandez,
MD MPH4, Christopher O’Connor MD
8, Randall C Starling MD MPH
9,
and W.H. Wilson Tang MD9
From the 1
Division of Cardiology, Department of Internal Medicine, University of Texas
Southwestern Medical Center, Dallas, TX ([email protected]); 2Department of
Medicine, University of Mississippi, Jackson, MS ([email protected]); 3Department of
Cardiology, University of Brescia, Brescia, Italy ([email protected]); 4Duke University
Medical Center, Duke Clinical Research Institute, Durham, NC ([email protected];
[email protected]); 5Hanzeplein 1, University Med Center Groningen, Groningen,
Netherlands ([email protected]); 6University of Glasgow, Glasgow, United Kingdom
([email protected]); 7Department of Cardiology, University of Alberta, Edmonton,
Canada ([email protected]); 8Inova Heart & Vascular Institute, Falls Church, VA
([email protected]); and 9Heart and Vascular Institute, Cleveland Clinic,
Cleveland, OH ([email protected]; [email protected]).
Address for Correspondence:
W. H. Wilson Tang, MD
9500 Euclid Avenue, Desk J3-4, Cleveland, OH 44195.
Phone: (216) 444-2121 / Fax: (216) 445-6165
E-mail: [email protected]
3
ABSTRACT
BACKGROUND. Circulating cardiac troponin levels (cTn), representative of myocardial injury,
are commonly elevated in heart failure (HF) and related to adverse clinical events. However,
whether cTn represents a spectrum of risk in HF is unclear.
METHODS. Baseline, 48–72 hour, and 30 day plasma cTnI was measured by a novel highly-
sensitive assay in 900 subjects with acute decompensated HF (ADHF) in ASCEND-HF.
Multivariable models determined the relationship between cTnI and outcomes.
RESULTS. The median(interquartile range) cTnI was 16.4 (9.3-31.6) ng/L at baseline, 14.1
(7.8-29.7) ng/L at 48-72 hours, and 11.6 (6.8-22.5) ng/L at 30 days. After additional adjustment
for amino terminal pro-B-type natriuretic peptide (NT-proBNP) to established risk predictors,
both baseline and 48-72 hour cTnI were associated with higher risk for death or worsening HF
prior to discharge (OR 1.25, P=0.03 and OR 1.43, P=0.001, respectively). However, only cTnI at
30 days was associated 180-day death (HR 1.25, P=0.007). There were no curvilinear
associations between changing cTnI and clinical outcomes.
CONCLUSIONS. Circulating cTnI level was associated with clinical outcomes in ADHF, but
these observations diminished with additional adjustment for NT-proBNP. Although they likely
represent a spectrum of risk in ADHF, these findings question the implications of changing cTnI
levels during treatment.
KEY WORDS. Troponin, nesiritide, acute decompensated heart failure
4
LIST OF ABBREVIATIONS:
cTnI – Cardiac troponin I
ADHF – Acute decompensated heart failure
cTn – Cardiac troponin
cTnT – Cardiac troponin T
URL – upper reference limit
ASCEND-HF - Acute Study of Clinical Effectiveness of Nesiritide in Decompensated Heart
Failure
NRI - net reclassification improvement
IDI - integrated discrimination improvement
NT-proBNP – amino terminal pro-B-type natriuretic peptide
BUN – blood urea nitrogen
LVEF – left ventricular ejection fraction
5
INTRODUCTION
Circulating cardiac troponin levels are commonly elevated in patients with heart failure. These
elevations likely represent a diversity of adverse myocardial responses to hemodynamic stress
and abnormalities in cardiomyocyte pathobiology.1 It is therefore no surprise that there is an
increasing interest in the clinical usage of circulating cTn in patients with heart failure. Over the
past decade, multiple epidemiologic studies have investigated both the prognostic role and
clinical implications of static or changing cTn levels in heart failure.2-12
Even though cTn levels
were measured in these studies by assays with widely varying sensitivities and cTn subtypes
(cTnT and cTnI), there is general agreement that higher circulating cTn is associated with worse
outcomes across the spectrum of HF patients. Unlike cTn in acute coronary syndromes where
cTn levels >99% URL, determined by each assay, have a high clinical sensitivity for detecting a
myocardial infarction,13
there is uncertainty as to whether there is an appropriate cTn cut-off for
risk in heart failure. Although, circulating cTn levels below the 99% URL in patients with acute
HF may be of low, long-term risk.14
Yet, there remains a prognostic role for very low levels of
circulating cTn2, especially when only detectable by increasingly sensitive assays,
15 and whether
there is a prognostic advantage apart from lower cTn detection is not clear. Unlike myocardial
infarction with a clearer link to pathology and cTn level, changes in cTn levels in HF are not
well understood.
The ASCEND-HF biomarker substudy provides a unique opportunity to answer these
questions.16
We aim to characterize the relationship of cTnI levels measured by a novel highly
sensitive assay with short-term, intermediate, and long-term endpoints and make a comparison to
cTnI levels measured by a concomitant standard sensitivity assay.11
Also, we aim to determine
6
the impact of nesiritide on serial cTnI levels and provide a better understanding of changing cTnI
levels in ADHF.
MATERIALS AND METHODS
STUDY POPULATION. The ASCEND-HF trial was a multicenter, randomized, double-blind
placebo-controlled trial that sought to determine the effects of nesiritide, a recombinant B-type
natriuretic peptide with vasodilating properties, in patients hospitalized with ADHF. Participants
with clinical evidence for an acute coronary syndrome or baseline cardiac troponin >5x the upper
reference limit, as measured by the local clinical laboratory, were excluded. The design and
primary results of ASCEND-HF have been previously reported 16,17
. Of the 7,411 patients
randomly assigned, 904 were enrolled in the biomarker sub-study. This is not a validation cohort,
and overall the sub-study population is comparable to that of the ASCEND-HF study population
(Supplemental Table 1). These subjects had serial venous blood sampling at baseline, 48–72
hours after therapy initiation, and at a 30-day follow-up visit 11,18
. Blood samples were collected
in ethylenediaminetetraacetic acid–plasma, immediately centrifuged and stored at -80ºF for
subsequent analysis at a core laboratory.
CARDIAC TROPONIN I MEASUREMENT. Circulating plasma cardiac troponin I levels
were determined in 900 of these individuals with a novel highly sensitive Single Molecule
Counting immunoassay technology which has been previously described19
. Briefly, cTnI levels
in 50 µl plasma samples were quantified using a plate-based sandwich immunoassay. At the
completion of the assay, single molecules of detection antibody were counted on the Erenna®
System (Singulex). The reporting range spanned 0.4–600 ng/nL, with inter-assay coefficient of
variation ranging from 0–20% with a 99th
percentile URL of 7.1 ng/L. Samples reporting results
7
>600 ng/L, were diluted 1:10 and retested to obtain a result within the assay reporting range.
Circulating plasma cTnI was also measured with a contemporary sensitive assay (VITROS Trop
I ES, Ortho Clinical Diagnostics, Raritan, NJ, USA) with a lower limit of detection of 0.012
ng/mL, a 99th
percentile URL of 0.034 ng/mL, and a coefficient of variation of 10% at the 99th
percentile URL.
CLINICAL ENDPOINTS. The co-primary endpoints from the ASCEND-HF trial were
dyspnea improvement (via 7-point Likert scale) at 6 to 24 hours and the composite endpoint of
death or recurrent heart failure hospitalization within 30 days of randomization. Death and
worsening heart failure were assessed together as a composite secondary endpoint and all events
were adjudicated through 180 days.
STATISTICAL ANALYSES. Baseline characteristics and clinical endpoints (for both placebo
and nesiritide arms) were presented as mean ± standard deviation or median (interquartile range),
where appropriate, for continuous variables and as count (percentage) for categorical variables.
Right-skewed variables were transformed via log base 2 so that an increase in one unit
represented adoubling of the variable values. The Jonckheere-Terpstra and Cochran-Armitage
trend tests were used to assess the significance of a trend across increasing tertiles of cTnI for
continuous and categorical variables, respectively. Two-sided p-values <0.05 were considered
statistically significant. Circulating cTnI levels were compared between patients randomly
assigned to placebo or nesiritide via the Wilcoxon rank sum test. In a hypothesis-generating
analysis, multivariable predictors of baseline and percent change in cTnI levels (modeled as a
spline functions for either an increase or decrease in cTnI from baseline) were determined via
backwards elimination linear regression modelling from candidate variables collected which
included demographic, clinical, and metabolic characteristics (Supplemental Table 1). In this
8
analysis, variables with P>0.1 were excluded from the model and P-values were bias-corrected
from 500 bootstrapped samples. Kaplan-Meier failure estimates according to baseline cTnI tertile
were compared by the Log-rank test. The association between cTnI and intermediate-term
outcomes was performed via logistic regression analyses for death or worsening heart failure in
the hospital, and 30-day death or hospitalization for worsening heart failure. The association of
cTnI and the long-term outcome 180-day death was performed via Cox-proportional hazards
models. The proportional hazards assumption was verified by determining if there were trends
with time for the Schoenfeld residuals. Fractional polynomial curves were generated to
demonstrate the continuous multivariable risk-adjusted probability of clinical outcomes across
cTnI levels. For multivariable analyses, we adjusted for covariates that have been identified for
the overall ASCEND-HF study population, as previously described 11,18
. Differences in 180-day
death curves determined by the log-rank test. The discrimination (C-statistic) of the impact of
cTnI levels on 180-day death was determined as previously described.20
Category-free NRI and
IDI were calculated for the addition of cTnI to the ASCEND-HF model.21
All statistical analyses
were performed using Stata software, version 13.1 (StataCorp LP, College Station, TX).
RESULTS
STUDY POPULATION. Of the 900 patients with baseline cTnI levels, 723 had cTnI levels
drawn at 48-72 hours, and 599 had cTnI levels drawn at 30 days. The median circulating cTnI
level at baseline was 16.4 (9.3, 31.6) ng/L. The median cTnI level at 48-72 hours was 14.1 (7.8,
29.7) ng/L and at 30 days was 11.6 (6.8, 22.5) ng/L. There were no patients below the lower
limit of quantification at baseline or at 30-days, and except one patient (0.14%); no others at 48-
72 hours were below the lower limit of quantification. Figure 1 illustrates the relationship of
9
measured cTnI by both the highly sensitive and standard sensitivity assays. Baseline
characteristics stratified by baseline circulating cTnI levels of this substudy cohort are shown in
Table 1. Higher circulating cTnI levels were associated with male sex, less prevalent white race,
lower left ventricular ejection fraction (LVEF), and less prevalent diabetes mellitus. It was also
associated with higher N-terminal prohormone brain natriuretic peptide (NT-proBNP) levels,
higher blood urea nitrogen (BUN) levels, and lower estimated glomerular filtration rate (eGFR).
SERIAL cTnI LEVELS. Table 2 demonstrates that there were no associations between
nesiritide randomization and change in troponin level from either baseline to 48-72 hours or 30
days.
DEMOGRAPHIC, CLINICAL, AND METABOLIC PREDICTORS OF cTnI LEVELS.
The complete results from backwards selection are shown in the table within Supplemental Table
1. Significant predictors of baseline cTnI included hs-CRP (P=0.049) and the LVEF above
≥40% (P=0.047), with the strongest associations with African American race (P=0.008), BUN
(P=0.004), and NT-proBNP (P<0.001). Interestingly, there was a strong inverse association with
diabetes (P=0.008). Significant predictors of increasing cTnI from baseline to 48-72 hours were
age below 55 years (P=0.007), potassium (P=0.019), and body mass index >30 kg/m2 (P=0.038).
On the other hand, significant predictors of decreasing cTnI from baseline to 48-72 hours were
baseline cTnI level (P<0.001), per 250 pg/mL NT-proBNP change from baseline to 48-72 hours
(P<0.001), ischemic heart failure (P=0.004); cerebrovascular disease (P=0.006); NT-proBNP
(P=0.015); and hs-CRP (P=0.015).
ASSOCIATION OF BASELINE AND FOLLOW-UP cTnI AND OUTCOMES. There were
42 (4.7%) deaths or worsening heart failure events by discharge. ,105 deaths or recurrent heart
failure hospitalizations and by 30 days, and 106 deaths (11.9%) by 180 days. Cumulative event
10
rates for these two outcomes according to baseline cTnI are shown in Figure 2. All univariable
and multivariable models for clinical outcomes are shown in Table 3. In univariable analyses,
higher baseline, 48-72 hour, and 30-day log base 2-transformed cTnI was associated with all
adverse clinical outcomes (P<0.01 for all) but not baseline log base 2-transformed cTnI and
death/rehospitalization at 30 days (P=0.06). Baseline log base 2-transformed cTnI was associated
with death or worsening heart failure prior to discharge (P=0.02) and with 180-day death
(P=0.02) after multivariable adjustment for the variables in Model 1. However, baseline log base
2-transformed cTnI only remained associated with death or worsening heart failure prior to
discharge after additional adjustment for NTproBNP (P=0.03). Log base 2-transformed cTnI at
48-72 hours and at 30-days was associated with all outcomes after multivariable adjustment for
the variables in Model 1 (P<0.05 for all). However, with additional adjustment for NTproBNP,
48-72 hours log base 2-transformed cTnI only remained associated with death or worsening heart
failure prior to discharge (P=0.001). The continuous, risk-adjusted (Model 2) curvilinear
associations between log base 2-transformed cTnI at baseline and 48-72 hours and % change in
cTnI from baseline to 48-72 hours with clinical outcomes is shown in Figure 3. While these
curves depict an increase in risk with increasing log base 2-transformed cTnI when measured
statically at baseline and 48-72 hours, the relative difference between the two values did not
appear to have a relationship with any of the clinical outcomes.
In a separate analysis (Supplemental Table 3) in subjects in whom cTnI was measured by
both the highly sensitive and standard sensitivity assays, the performance of the ASCEND-HF
models with cTnI added was mildly improved. Both assays measured at the 3 time points
performed comparably in terms of improvement in model fit, yet there was no added model
discrimination (C-statistics) or integrated discrimination improvement. Only the highly sensitive
11
assay had significant net reclassification for death or rehospitalization by 30 days (NRI 36.3%,
P=0.007) and death or worsening heart failure prior to discharge (NRI 45.6%, P=0.02). The
prognostic relevance of the standard sensitivity cTnI assay in this biomarker substudy of
ASCEND-HF has been previously published.11
DISCUSSION
This analysis from ASCEND-HF highlights several new insights regarding the utility of
circulating cTnI levels in patients with ADHF. First, cTnI was detectable in all but one subject at
all three time points, with the majority having levels >99% URL. Second, compared with
placebo, nesiritide was not associated with changing cTnI levels. Third, only neurohormonal
biomarkers (NTproBNP and BUN) and few demographic variables (African American Race and
Diabetes) were significant predictors of cTnI level. Fourth, circulating cTnI levels were highly
predictive of both short-term and long-term adverse events, but might not be incremental to more
traditional ADHF markers of risk or hemodynamic stress. Interestingly, there appeared to be a
time-dependency with the time-point of cTnI measurement (baseline, 48-72 hours, and 30 days)
and the outcomes for which it was associated. In aggregate, these findings highlight that cTnI
levels represent a spectrum of risk in heart failure with the possibility of risk-stratification at
different time points. In particular, follow-up cTnI levels were more useful than baseline to
establish prognosis – findings which oppose the current heart failure treatment guideline's
emphasis on admission cTnI.22
Circulating cardiac troponin levels are rarely detectable in the general population, 23-26
but when in the measureable range they are associated with cardiac abnormalities and adverse
outcomes. In patients with heart failure, however, circulating cardiac troponin levels are higher
12
and more frequently detectable. There is a litany of studies showing an increased risk for adverse
outcomes with higher circulating cardiac troponin levels (measured by both standard and highly
sensitive assays) in both acute and chronic heart failure.2-10
While the precise reasons for this are
unclear, higher cardiac troponin may indicative of progressive myocardial dysfunction with its
release likely triggered by acute or chronic myocardial stress, sub-endocardial ischemia, or direct
cardiomyocyte injury.27
In a marked contrast to a comparable analysis detecting cTnI by a contemporary, but less
sensitive assay (LOD 0.012 ng/ml, which is 30x less sensitive) in the same ASCEND-HF
population where cTnI was detectable in 78% of the population with 50% having cTnI levels
above the 99% URL,11
we found cTnI was detectable in virtually all patients by this more highly
sensitive assay at all three time points (100%, 99.9%, and 100%, respectively) with the majority
above the 99% URL (83%, 78%, and 73%, respectively) and the relationship between these two
assays when measured together is shown in Figure 1. This striking difference is no doubt a result
of a more highly sensitive cTnI assay and is relatively consistent with a prior analysis from
RELAX-AHF, which suggested showed that cTnT measured by a highly sensitive assay was
above the 99% URL in 90% of patients with acute heart failure.12
Interestingly, neither cTnI
assay improved risk stratification when added to other strong risk factors in the ASCEND-HF
population. The implications of this in HF are twofold: 1) calling into question the clinical utility
of measuring cTnI changes; and 2) that, if in the measurable range, cTn may have little
interassay prognostic difference.
Regardless, this near-total prevalence of detectable circulating cTnI serves as a sobering
reminder that hemodynamic stress from ADHF is associated alterations in cardiomyocyte
homeostasis, even on the “ultra-small” scale; and that subtle differences in circulating cTnI, in
13
contrast to “positive” or “negative”, may represent a spectrum of processes within the
myocardium during decompensation or hemodynamic stress resulting in troponin release or
turnover. In addition, the lack of association between cTnI levels and nesiritide supports the
hypothesis that cTnI levels may change may be additionally mediated by non-hemodynamic
variables and questions its utility as a surrogate for safety or efficacy for other therapies that do
not directly impact cardiomyocytes.
Cardiac troponin release may be a result of the impact of several adverse pathological
processes on the myocardium, but its changes during the treatment of ADHF are less well
understood. Interestingly, our finding that both NT-proBNP and BUN are strong independent
predictors of circulating cTnI levels suggests that both myocardial wall stress and neurohormonal
activation may be related to troponin release. Natriuretic peptide elevation is a compensatory
response to the direct effects of hemodynamic stress the myocardium and is a biochemical
surrogate for elevated cardiac filling pressures;28,29
and blood urea nitrogen is commonly
recognized as surrogate for both glomerular and renal tubular function, but is closely correlated
to neurohormonal activation such as the heightened renin-angiotensin-aldosterone-system
activity in heart failure.30
Decreasing cTnI during treatment is most closely associated impacted
by baseline troponin release and is halted by a history of ischemic heart failure. However, the
opposite directionality of NT-proBNP and decreasing cTnI suggests that cTnI levels are
impacted from other non-hemodynamic factors.
There is a growing body of literature regarding the prognostic of circulating cTn levels in
heart failure populations. With the transition from earlier studies analyzing cTn by standard
sensitivity assays to the current more modern highly-sensitive assays in which cTn is nearly
ubiquitously detectable, circulating cTn likely represents a spectrum or risk for heart failure
14
populations. This relationship is well represented in Figure 3, as there is a graded increase in
adjusted risk for all outcomes for increasing cTnI levels. Unlike prior analyses that have shown
higher risk for increasing troponin levels in patients with ADHF, 12,31,32
we found no consistent
association with changing cTnI levels and outcomes. Although, the shape of curves suggests
higher risk for both a decrease in cTnI or increase in cTnI for 30-day adverse outcomes, perhaps
identifying a risk in intermediate events with cTnI level lability. While there was no association
with changing cTnI levels and nesiritide usage, cTn levels may be mediated by other therapies
such as seralaxin.32
STUDY LIMITATIONS. The results of this study must be interpreted in light of various
limitations inherent to its design. First, this analysis was post-hoc, which renders these findings
hypothesis generating. Second, we cannot exclude selection bias for patients whom agreed to
participate in the ASCEND-HF trial. Although, the baseline clinical characteristics (Table 1) of
this cohort are representative of contemporary populations hospitalized with ADHF with the
majority on guideline directed medical therapies and presenting with marked NT-proBNP
elevation. Third, even though nesiritide use was not associated with changes in circulating cTnI
levels, the impact of other therapies on cTnI or whether they modified the prognostic effect of
cTnI on clinical outcomes remains unclear. Regardless, with carefully adjudicated outcomes and
cTnI measured in a core laboratory, our analysis suggests that circulating cTnI levels are
independently prognostic in ADHF and, even when detected at low levels by this highly
sensitive assay, are associated with a spectrum of risk. We also acknowledge that prospective
studies in both ADHF and chronic heart failure (and by LVEF), within a multi-marker setting,
are necessary to establish the clinical importance of cTn measured by highly sensitive assays.
15
CONCLUSIONS
In our study cohort, circulating cTnI measured at baseline, during hospitalization, and at 30-days
was associated with adverse clinical outcomes in ADHF, but may not be incremental to more
traditional predictors of risk. Interestingly, after additional adjustment for NT-proBNP, higher
circulating cTnI was more closely associated with death or worsening heart failure prior to
discharge in comparison to outcomes occurring more remotely from discharge. Unlike prior
reports, these findings question the implications of changing cTnI levels in ADHF.
Sources of funding: The ASCEND-HF study, including the biomarker substudy, was funded by
Scios Inc.; Janssen Research & Development LLC retains operational responsibility for the
ASCEND-HF study. Singulex, Inc. performed all plasma cTnI assays, and was blinded from the
trial database or analyses. Statistical analyses, and manuscript preparation were conducted
independent of the sponsors, and the authors have access to all the data in its entirety and
approved the final manuscript.
16
REFERENCES
1. Kociol RD, Pang PS, Gheorghiade M, Fonarow GC, O'Connor CM, Felker GM.
Troponin elevation in heart failure prevalence, mechanisms, and clinical implications. J
Am Coll Cardiol. 2010;56(14):1071-1078.
2. Latini R, Masson S, Anand IS, et al. Prognostic value of very low plasma concentrations
of troponin T in patients with stable chronic heart failure. Circulation.
2007;116(11):1242-1249.
3. Masson S, Anand I, Favero C, et al. Serial measurement of cardiac troponin T using a
highly sensitive assay in patients with chronic heart failure: data from 2 large randomized
clinical trials. Circulation. 2012;125(2):280-288.
4. Nagarajan V, Hernandez AV, Tang WH. Prognostic value of cardiac troponin in chronic
stable heart failure: a systematic review. Heart. 2012;98(24):1778-1786.
5. Peacock WFt, De Marco T, Fonarow GC, et al. Cardiac troponin and outcome in acute
heart failure. N Engl J Med. 2008;358(20):2117-2126.
6. Metra M, Nodari S, Parrinello G, et al. The role of plasma biomarkers in acute heart
failure. Serial changes and independent prognostic value of NT-proBNP and cardiac
troponin-T. Eur J Heart Fail. 2007;9(8):776-786.
7. You JJ, Austin PC, Alter DA, Ko DT, Tu JV. Relation between cardiac troponin I and
mortality in acute decompensated heart failure. Am Heart J. 2007;153(4):462-470.
8. Xue Y, Clopton P, Peacock WF, Maisel AS. Serial changes in high-sensitive troponin I
predict outcome in patients with decompensated heart failure. Eur J Heart Fail.
2011;13(1):37-42.
17
9. Perna ER, Macin SM, Cimbaro Canella JP, et al. Importance of early combined N-
terminal pro-brain natriuretic peptide and cardiac troponin T measurements for long-term
risk stratification of patients with decompensated heart failure. J Heart Lung Transplant.
2006;25(10):1230-1240.
10. Pascual-Figal DA, Manzano-Fernandez S, Boronat M, et al. Soluble ST2, high-sensitivity
troponin T- and N-terminal pro-B-type natriuretic peptide: complementary role for risk
stratification in acutely decompensated heart failure. Eur J Heart Fail. 2011;13(7):718-
725.
11. Felker GM, Hasselblad V, Tang WH, et al. Troponin I in acute decompensated heart
failure: insights from the ASCEND-HF study. Eur J Heart Fail. 2012;14(11):1257-1264.
12. Felker GM, Mentz RJ, Teerlink JR, et al. Serial high sensitivity cardiac troponin T
measurement in acute heart failure: insights from the RELAX-AHF study. Eur J Heart
Fail. 2015;17(12):1262-1270.
13. Thygesen K, Alpert JS, Jaffe AS, et al. Third universal definition of myocardial
infarction. J Am Coll Cardiol. 2012;60(16):1581-1598.
14. Pang PS, Teerlink JR, Voors AA, et al. Use of High-Sensitivity Troponin T to Identify
Patients With Acute Heart Failure at Lower Risk for Adverse Outcomes: An Exploratory
Analysis From the RELAX-AHF Trial. JACC Heart Fail. 2016;4(7):591-599.
15. Grodin JL, Neale S, Wu Y, Hazen SL, Tang WH. Prognostic comparison of different
sensitivity cardiac troponin assays in stable heart failure. Am J Med. 2015;128(3):276-
282.
16. O'Connor CM, Starling RC, Hernandez AF, et al. Effect of nesiritide in patients with
acute decompensated heart failure. N Engl J Med. 2011;365(1):32-43.
18
17. Hernandez AF, O'Connor CM, Starling RC, et al. Rationale and design of the Acute
Study of Clinical Effectiveness of Nesiritide in Decompensated Heart Failure Trial
(ASCEND-HF). Am Heart J. 2009;157(2):271-277.
18. Tang WH, Dupont M, Hernandez AF, et al. Comparative assessment of short-term
adverse events in acute heart failure with cystatin C and other estimates of renal function:
results from the ASCEND-HF trial. JACC Heart Fail. 2015;3(1):40-49.
19. Todd J, Freese B, Lu A, et al. Ultrasensitive flow-based immunoassays using single-
molecule counting. Clin Chem. 2007;53(11):1990-1995.
20. Harrell FE, Jr., Lee KL, Mark DB. Multivariable prognostic models: issues in developing
models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat
Med. 1996;15(4):361-387.
21. Pencina MJ, D'Agostino RB, Sr., Steyerberg EW. Extensions of net reclassification
improvement calculations to measure usefulness of new biomarkers. Stat Med.
2011;30(1):11-21.
22. Yancy CW, Jessup M, Bozkurt B, et al. 2017 ACC/AHA/HFSA Focused Update of the
2013 ACCF/AHA Guideline for the Management of Heart Failure: A Report of the
American College of Cardiology/American Heart Association Task Force on Clinical
Practice Guidelines and the Heart Failure Society of America. J Card Fail.
2017;23(8):628-651.
23. Wallace TW, Abdullah SM, Drazner MH, et al. Prevalence and determinants of troponin
T elevation in the general population. Circulation. 2006;113(16):1958-1965.
19
24. de Lemos JA, Drazner MH, Omland T, et al. Association of troponin T detected with a
highly sensitive assay and cardiac structure and mortality risk in the general population.
JAMA. 2010;304(22):2503-2512.
25. Blankenberg S, Salomaa V, Makarova N, et al. Troponin I and cardiovascular risk
prediction in the general population: the BiomarCaRE consortium. Eur Heart J.
2016;37(30):2428-2437.
26. Everett BM, Zeller T, Glynn RJ, Ridker PM, Blankenberg S. High-sensitivity cardiac
troponin I and B-type natriuretic Peptide as predictors of vascular events in primary
prevention: impact of statin therapy. Circulation. 2015;131(21):1851-1860.
27. Feng J, Schaus BJ, Fallavollita JA, Lee TC, Canty JM, Jr. Preload induces troponin I
degradation independently of myocardial ischemia. Circulation. 2001;103(16):2035-
2037.
28. Nakagawa K, Yamaguchi T, Seida M, et al. Plasma concentrations of brain natriuretic
peptide in patients with acute ischemic stroke. Cerebrovasc Dis. 2005;19(3):157-164.
29. Maeda K, Tsutamoto T, Wada A, Hisanaga T, Kinoshita M. Plasma brain natriuretic
peptide as a biochemical marker of high left ventricular end-diastolic pressure in patients
with symptomatic left ventricular dysfunction. Am Heart J. 1998;135(5 Pt 1):825-832.
30. Schrier RW. Blood urea nitrogen and serum creatinine: not married in heart failure. Circ
Heart Fail. 2008;1(1):2-5.
31. O'Connor CM, Fiuzat M, Lombardi C, et al. Impact of serial troponin release on
outcomes in patients with acute heart failure: analysis from the PROTECT pilot study.
Circ Heart Fail. 2011;4(6):724-732.
20
32. Metra M, Cotter G, Davison BA, et al. Effect of serelaxin on cardiac, renal, and hepatic
biomarkers in the Relaxin in Acute Heart Failure (RELAX-AHF) development program:
correlation with outcomes. J Am Coll Cardiol. 2013;61(2):196-206.
21
Table 1. Baseline characteristics
cTnI [ng/L]
Total Tertile 1, < 11.3 Tertile 2, 11.3 - 24.0 Tertile 3, > 24.0 *P-value
(N=900) (N=300) (N=300) (N=300)
Age [years] 65.6 +/- 15.1 64.8 +/- 15.6 66.4 +/- 14.3 65.7 +/- 15.4 0.53
Sex (female, %) 284 (31.6) 125 (41.7) 85 (28.3) 74 (24.7) <.001
Race (white, %) 610 (67.8) 212 (70.7) 209 (69.7) 189 (63.0) 0.045
BMI [kg/m²] 32.0 +/- 8.8 32.8 +/- 9.0 31.0 +/- 7.6 32.1 +/- 9.7 0.31
LVEF [%] 30.7 +/- 15.4 33.6 +/- 17.0 28.6 +/- 14.4 29.9 +/- 14.4 0.006
NYHA Classification
I 33 (5.0) 10 (4.5) 13 (6.2) 10 (4.4)
II 147 (22.3) 47 (21.1) 45 (21.4) 55 (24.4)
III 318 (48.3) 113 (50.7) 97 (46.2) 108 (48.0)
IV 160 (24.3) 53 (23.8) 55(26.2) 52 (23.1)
Diabetes Mellitus 414 (46.0) 143 (47.7) 156 (52.0) 115 (38.3) 0.022
Hypertension 704 (78.2) 234 (78.0) 241 (80.3) 229 (76.3) 0.62
Ischemic Heart Failure 540 (60.0) 172 (57.3) 190 (63.3) 178 (59.3) 0.62
Beta-Blockers 676 (75.1) 228 (76.0) 233 (77.7) 215 (71.7) 0.22
ACEI or ARB 579 (64.3) 193 (64.3) 193 (64.3) 193 (64.3) 1
MRA 219 (24.3) 77 (25.7) 76 (25.3) 66 (22.0) 0.30
Time-to-randomization
[hours] 16.4 +/- 9.4 16.9 +/- 9.2 16.2 +/- 10.3 16.0 +/-8.6 0.29
NT-proBNP [pg/mL] 5791 (2986-11579) 4073 (2144-7444) 7499 (3498-12920) 7582 (3583-12984) <.001
BUN [mmol/L] 10.3 +/- 6.1 8.8 +/- 5.0 10.6 +/- 6.0 11.4 +/- 6.9 <.001
eGFR [mL/min/1.73m²] 59.8 +/- 24.5 64.8 +/- 24.9 56.5 +/- 23.2 58.0 +/- 24.8 0.001
Hemoglobin [g/dL] 12.6 +/- 4.9 12.9 +/- 1.8 12.5 +/- 2.0 12.6 +/- 1.9 0.45
Sodium [mmol/L] 138.5 +/- 3.9 138.1 +/- 4.1 138.6 +/- 3.8 138.9 +/- 3.9 0.011
*p-value for trend
22
Table 2. Effect of Nesiritide Therapy on Serial Changes in cTnI Levels*
cTnI Levels [ng/L] Placebo (N=450) Nesiritide (N=450) P-value
Baseline 15.9 (8.7, 29.2) 16.9 (9.5, 34.1) 0.39
Actual change from baseline to 48–72 hours -2.1 (-7.1, 1.1) -1.7 (-6.9, 1.0) 0.78
Actual change from baseline to 30 days -2.3 (-11.1, 1.9) -2.2 (-10.0, 1.6) 0.65
Relative change from baseline to 48–72 hours [%] -18.4 (-40.2, 11.3) -15.2 (-34.9, 8.9) 0.60
Relative change from baseline to 30 days [%] -22.2 (-52.2, 16.0) -18.2 (-49.6, 19.2) 0.46
*Variables expressed as median (Q1-Q3)
23
Table 3. cTnI Levels and in-hospital-, intermediate-, and long-term outcomes
Endpoint *Unadjusted *Adjusted Model 1 *Adjusted Model 2
†OR/HR 95% CI P-Value †OR/HR 95% CI P-value †OR/HR 95% CI P-value
Baseline cTnI Level
Death or worsening heart failure prior
to discharge
1.28 1.09-1.50 0.003 1.26 1.04-1.52 0.02 1.25 1.02-1.55 0.03
Death or HF hospitalization at 30 Days 1.12 0.997-1.27 0.06 1.07 0.94-1.22 0.30 1.07 0.93-1.22 0.37
Death at 180 days 1.18 1.06-1.31 0.002 1.14 1.02-1.26 0.02 1.09 0.97-1.22 0.16
48-72 Hour cTnI Level
Death or worsening heart failure prior
to discharge
1.42 1.20-1.67 <.001 1.42 1.17-1.73 0.001 1.43 1.17-1.75 0.001
Death or HF hospitalization at 30 Days 1.20 1.06-1.35 0.003 1.14 1.004-1.30 0.04 1.12 0.98-1.28 0.08
Death at 180 days 1.17 1.05-1.31 0.006 1.16 1.03-1.31 0.02 1.11 0.97-1.26 0.14
30 day cTnI Level
Death at 180 days 1.23 1.08-1.40 0.001 1.29 1.11-1.50 0.001 1.25 1.06-1.47 0.007
*Unadjusted: only natural log base 2-transformed cTnI; Adjusted Model 1 = log base 2-transformed cTnI adjusted with covariates derived from the original
ASCEND-HF population; andAdjusted Model 2 = log base 2 -transformed cTnI adjusted with covariates derived from the original ASCEND-HF population
including natural log base 2-transformed NT-proBNP.
†Cox proportional hazards models (HRs) determined the association between log base 2-transformed cTnI and death at 180 days while logistic regression
24
models (ORs) were used for all other outcomes.
25
FIGURES
Figure 1.
Title: Distribution of cardiac troponin I levels measured by the highly sensitive assay.
Caption: All values are > the lower limit of quantification (<0.04 ng/L) for the highly sensitive
cardiac troponin I assay. For the cardiac troponin I levels measured by the standard assay; 170
(21.5%) subjects were undetectable (<0.012 ng/mL); 225 (28.1%) subjects were detectable, but
below the 99th
percentile URL (≥0.012 ng/mL to ≤0.034 ng/mL); and 396 (50.1%) subjects were
above the 99th
percentile URL (>0.034 ng/mL). URL, upper reference limit.
26
Figure 2.
Title: Kaplan-Meier estimates of intermediate- and long-term outcomes according to baseline
cTnI tertile
Caption: A) 30-day death or HF hospitalization and B) 180-day death.
27
Figure 3.
Title: The risk-adjusted, continuous probability of death or worsening HF by discharge, 30-day
death or HF hospitalization, and 180-day death for baseline cTnI, 48-72 hour cTnI, and percent
change in cTnI from baseline to 48-72 hours.
Caption: All models are risk-adjusted with covariates derived from the original ASCEND-HF
population with the addition of NT-proBNP. The vertical axes represent predicted probabilities.
For the baseline and 48-72 hour cTnI, the horizontal axes represent log base 2-transformed
values (every doubling).